in the Wadden Sea (Schiermonnikoog, The Nether-
lands; Terschelling, The Netherlands; Skallingen,
Denmark), this appeared to be a general phenome-
non (Olffet al.1997; Van Wijnen and Bakker 1997).
Soil nitrogen is a limiting factor for plant produc-
tion in salt-marsh systems (see overview in Davyet al.
2009). As the nitrogen availability is positively related
to the nitrogen pool (Bakkeret al. 2005b), the plant
productivity increases with a growing thickness of
clay layer (Van de Koppelet al. 1996). In other
words the chronosequence of increasing thickness of
sediment represents a productivity gradient.
10.3 Effects of plants on herbivores (bottom-up control)
Along the productivity gradient the density of
the wild herbivores such as different species
of Arctic geese (e.g. brent gooseBranta bernicla berni-
cla,barnacle gooseBranta leucopsis), brown hares
(Lepus europaeus) and rabbits (Oryctolagus cuniculus)
initially increases to anoptimum at intermediate pro-
ductivity, but declines at sites with high productivity
(Van de Koppelet al. 1996). According to theory, at
sites with low productivity, plant biomass is too low
to support a herbivore population, and plant growth
will be regulated by bottom-up effects such as nutri-
ent availability (Oksanen and Oksanen 2000). With
increasing productivity a shift from bottom-up to top-
down effects is expected to occur. Top-down regula-
tion of plant biomass occurs at sites of intermediate
levels of productivity, and herbivore population will
be top-down regulated by carnivores at high produc-
tivity (Oksanen and Oksanen 2000). However, in the
absence of carnivores (e.g. one of our study systems,
Schiermonnikoog) bottom-up effects remain to play
an important role even at highly productive sites.
Herbivore density can decrease even in the absence
of carnivores. Intake rate of geese levels off or declines
with biomass above a certain threshold (Van der
Graafet al. 2006). Forage quality declines at sites
of high biomass and tall canopy (Van der Walet al.
2000a; Kuijper and Bakker 2005) featuring a decreas-
ing leaf–stem ratio. This bottom-up control of herbi-
vore density at high productivity sites is referred to as
the ‘quality threshold hypothesis’ (Van de Koppelet
al. 1996; Olffet al. 1997; Huismanet al. 1999).
The productivity gradient (chronosequence) on
Schiermonnikoog is accompanied by plant species
replacement. The unproductive lower salt marsh
is dominated bySalicorniaspp.,Puccinellia maritima,
Plantago maritimaandLimonium vulgare, whereas
the oldest stages are dominated byAtriplex portula-
coides. The unproductive higher marsh features
Puccinellia maritimaandFestuca rubrafollowed by
Artemisia maritimaand, eventually,Elymus athericus
(Elytrigia atherica) at the productive marsh (Olffet
al. 1997; Van der Walet al. 2000a). Both at the low
and high salt marsh, succession eventually features
a tall canopy ofAtriplex portulacoidesorElymus
athericus, respectively. Recently, it was noticed
thatElymus athericusspread into lower elevation at
older marshes (Olffet al. 1997). These tall plant
species outcompete other species by light intercep-
tion (Huismanet al. 1999; Van der Walet al. 2000a),
with subsequent decline in plant species richness
(Bakkeret al. 2003b).
Herbivores are evicted by plant succession.
Goose numbers were estimated at young, interme-
diate and older parts of the salt marsh on Schier-
monnikoog between 1971 and 1997 (Fig. 10.2). In
the late 1970s brent goose numbers were high in the
old marsh. However, goose numbers declined sig-
nificantly in the following 20 years (Van der Walet
al. 2000b). This decrease is not related to a decrease
in size of the area. On the contrary, the surface area
increased over the years as a result of sedimenta-
tion (Fig. 10.1). Goose numbers increased in the
intermediate aged salt marsh followed by a slight
but significant decrease towards 1997. Develop-
ment of new young marsh in the east led to a
further eastward movement and an increase of
goose abundance (Van der Walet al. 2000b). The
decrease in number of brent geese at the older
marsh coincided with a change in vegetation com-
position. In 1977, when goose abundance was
still high, the clonal shrubAtriplex portulacoides
was lacking. Since then, theAtriplexcommunity
hasspread into the lower elevation salt marsh,
and this coincided with the observed decline in
goose numbers. Part of theLimoniumcommunity
was transformed into theAtriplex portulacoidescom-
munity. The openLimonium vulgarecommunity
harbours the preferred goose food plants such as
Puccinellia maritima, Festuca rubraandTriglochinCOMMUNITY ECOLOGY AND MANAGEMENT OF SALT MARSHES 133